THE EARTHQUAKE ADVANCE THEORY 189
ferent bergs. The white bergs owe their color to the presence of great numbers of included air bubbles, all of small size, and often very closely spaced; the glassy bergs have fewer air bubbles and sometimes none at all. In general the white bergs are the more finely-crystalline, though many of them are coarsely-crystalline; but the glassy bergs are always coarsely-crystalline. We assume that the air bubbles of the white bergs represent air imprisoned in the course of snow accumulation in the upper portions of the glacier and included in the ice during its crystallization.
Why is there this difference in the ice from the same glaciers? The obvious answer is that it is due to position in the glacier, the glassy, coarsely-crystalline bergs being from the lower ice, the more finely crystalline, air-charged, white ice from the upper portion; but in those frontal portions of glaciers where ablation has removed much of the upper ice, coarsely-crystalline ice forms the surface., This conclusion is supported by two facts: (1) the part of the glacier front above the water in the ice cliffs is white; (2) although some white bergs contain debris, the most heavily charged icebergs are invariably glassy. The problem then resolves itself into the question, why is the upper ice charged with air bubbles while in the lower ice there is no such abundance of air bubbles? We have tried to conceive of conditions under which this difference in ice condition could be brought about. It is inconceivable that the lower ice originally contained less imprisoned air than the upper ice, for it is from the same source of snowfall, compacted and consolidated into the crystalline ice. We are forced to conclude, therefore, that the air originally included in the lower region of glassy ice has been eliminated in the course of development of the coarsely-crystalline structure. That such elimination has been in progress within the glacier is indicated by the fact that air bubbles are present in great abundance in some coarsely-crystalline ice and in all the finely-crystalline bergs. By this fact we are forced to the conclusion that in some parts of the ice there are unusual conditions which permit the elimination of included air bubbles and that these conditions prevail mainly, if not entirely, in the bottom layers of the glacier.
From the consideration of the problem which we have been able to give we can find but one explanation that is capable of accounting for these facts. We accordingly propose as an hypothesis that the lower portion of the glacier is in a condition of sufficient viscosity to force the escape of the included air bubbles, which, rising, become included in the upper area of viscosity, giving rise to that layer of ice from which some of the whitest icebergs are derived. Whether the coarsely-crystalline structure is present in the viscous ice, or whether it develops as the front of the glacier is neared, and the pressure relieved, is not indicated by any facts which we have observed.
From the above statement of facts, and from the consideration of the inferences which seem to be warranted from them, we are led from various points of view to the same result. We are wholly unable to understand the phenomena observed in the Yakutat Bay region without resort to the theory of viscous flowage. By this theory all phenomena are readily explained, while by any other theory we find ourselves confronted by grave difficulties in accounting for some of the phenomena. We are, therefore, forced to believe that the theory of viscous ice motion has foundation, and that the phenomena of the advancing glaciers of Yakutat Bay are an expression of an unusually active phase of such flowage.
Lack of Similar Advance Elsewhere. Were it not for the spectacular nature of the